Make real time trascription like Assemble AI (only full sentence)

I wan't to achieve real time transcription like Assemble AI, (only full sentence)
https://www.youtube.com/watch?v=5LJFK7eOC20

I can be achieved by seamless by META
https://huggingface.co/spaces/facebook/seamless_m4t

https://github.com/facebookresearch/seamless_communication/tree/main

or sanchit-gandhi/whisper-jax by OpenAI and huggingface

https://huggingface.co/spaces/sanchit-gandhi/whisper-jax?fbclid=IwAR0fs6gXG8lBDFuuKGWqGjpf04QtF9v-Lqq5XvSWiS7FX1eDtQfJ4SAKJuw

https://github.com/sanchit-gandhi/whisper-jax

The most important for this is it has to be in real time so it has to listen to voice input and react fast and accurate.

So speed and accuracy are the main factors of success.
Please justify your bid how you could do that.
By choosing person for this project I will evaluate your experience, price and your bid comment.